Infectious Disease Prediction and Training Method, Device, Equipment and Medium

Abstract

The present application discloses an infectious disease prediction and training method, device, equipment and medium, belonging to the field of artificial intelligence. The prediction method includes: obtaining first physiological information and first location information of a first user; adding a first user node to a basic graph network of an infectious disease based on the first physiological information and the first location information to obtain an updated graph network; the basic graph network includes at least one user node and / or at least one area node, the user node identifies a user suspected of having an infectious disease, and the area node identifies the area where the user suspected of having an infectious disease is located; inputting the updated graph network into a graph convolutional neural network to predict a first user feature vector of the first user; and based on the first user feature vector, predicting the probability that the first user has an infectious disease through a probability calculation network. The above solution improves the accuracy of infectious disease prediction.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and particularly to an infectious disease prediction and training method, device, equipment and medium. Background Art

[0002] In the medical field, infectious disease prediction plays a crucial role in the prevention and control of infectious diseases.

[0003] In the related art, the probability that a suspected patient has a target infectious disease is predicted by determining whether the symptoms of the suspected patient are the symptoms of the target infectious disease; for example, if the suspected patient shows cough symptoms (symptoms of the target infectious disease), the related art marks the suspected patient as a high-risk person.

[0004] However, the related art cannot distinguish other similar diseases through symptom information. For example, the related art may mark a patient with a cold (cough is also a symptom of a cold) as a high-risk person with the target infectious disease, and the accuracy of infectious disease prediction by the related art is low. Summary of the Invention

[0005] The present application provides an infectious disease prediction and training method, device, equipment and medium, which can improve the accuracy of infectious disease prediction. The technical solutions are as follows:

[0006] According to one aspect of the present application, an infectious disease prediction method is provided, and the method includes:

[0007] Obtain the first physiological information and the first location information of a first user, where the first physiological information is the physiological information related to the infectious disease of the first user, and the first location information indicates the location where the first user is located;

[0008] Based on the first physiological information and the first location information, add a first user node to the basic graph network of the infectious disease to obtain an updated graph network; the basic graph network includes at least one user node and / or at least one area node, the user node identifies a user suspected of having an infectious disease, and the area node identifies the area where the user suspected of having an infectious disease is located;

[0009] Input the updated graph network into a graph convolutional neural network, and predict to obtain a first user feature vector of the first user; and, based on the first user feature vector, predict the probability that the first user has an infectious disease through a probability calculation network.

[0010] According to another aspect of the present application, a training method for a graph convolutional neural network is provided, and the method adopts a semi-supervised mode, and the method includes:

[0011] Obtain the sample physiological information and sample location information of n sample users; the n sample users include n1 first sample users and n2 second sample users, the first sample users carry labels, and the labels are used to determine whether the first sample users have infectious diseases, and the second sample users do not carry labels;

[0012] Input the sample graph network corresponding to the sample physiological information and sample location information of n sample users into a graph convolutional neural network, and calculate n sample feature vectors of n sample users through the graph convolutional neural network;

[0013] For one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has an infectious disease through a probability calculation network;

[0014] Train the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0015] According to another aspect of the present application, there is provided a method for training a graph convolutional neural network. The method adopts a supervised mode, and the method includes:

[0016] Obtain the sample physiological information and sample location information of n1 first sample users; the first sample users carry labels, and the labels are used to determine whether the first sample users have infectious diseases;

[0017] Input the sample graph network corresponding to the sample physiological information and sample location information of n1 first sample users into a graph convolutional neural network, and calculate n1 sample feature vectors of n1 first sample users through the graph convolutional neural network;

[0018] For one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has an infectious disease through a probability calculation network;

[0019] Train the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0020] According to another aspect of the present application, there is provided an infectious disease prediction device, and the device includes:

[0021] An acquisition module, configured to acquire the first physiological information and the first location information of a first user, where the first physiological information is the physiological information related to infectious diseases of the first user, and the first location information indicates the location where the first user is located;

[0022] An update module, configured to add a first user node to a basic graph network of infectious diseases based on first physiological information and first location information, to obtain an updated graph network; the basic graph network includes at least one user node and / or at least one region node, the user node identifies a user suspected of having an infectious disease, and the region node identifies the region where the user suspected of having an infectious disease is located;

[0023] A prediction module, configured to input the updated graph network into a graph convolutional neural network, and predict a first user feature vector of the first user; and, based on the first user feature vector, predict the probability that the first user has an infectious disease through a probability calculation network.

[0024] According to another aspect of the present application, there is provided a training device for a graph convolutional neural network, the device includes:

[0025] An acquisition module, configured to acquire sample physiological information and sample location information of n sample users; the n sample users include n1 first sample users and n2 second sample users, the first sample users carry labels, and the labels are used to determine whether the first sample users have infectious diseases, and the second sample users do not carry labels;

[0026] A calculation module, configured to input a sample graph network corresponding to the sample physiological information and sample location information of the n sample users into a graph convolutional neural network, and calculate n sample feature vectors of the n sample users through the graph convolutional neural network;

[0027] The calculation module is further configured to, for one of the n1 first sample users, obtain a predicted probability that the first sample user has an infectious disease based on the sample feature vector of the first sample user through a probability calculation network;

[0028] A training module, configured to train the graph convolutional neural network based on a loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0029] According to another aspect of the present application, there is provided a training device for a graph convolutional neural network, the device includes:

[0030] An acquisition module, configured to acquire sample physiological information and sample location information of n1 first sample users; the first sample users carry labels, and the labels are used to determine whether the first sample users have infectious diseases;

[0031] A calculation module, configured to input a sample graph network corresponding to the sample physiological information and sample location information of the n1 first sample users into a graph convolutional neural network, and calculate n1 sample feature vectors of the n1 first sample users through the graph convolutional neural network;

[0032] The calculation module is also used to obtain the predicted probability of a first-sample user having an infectious disease through a probability calculation network based on the sample feature vector of the first-sample user among n1 first-sample users.

[0033] The training module is used to train the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among n1 first-sample users.

[0034] According to one aspect of the present application, there is provided a computer device, which includes: a processor and a memory. The memory stores a computer program, and the computer program is loaded and executed by the processor to implement the above-mentioned infectious disease prediction method and / or the training method of the graph convolutional neural network.

[0035] According to another aspect of the present application, there is provided a computer-readable storage medium, which stores a computer program, and the computer program is loaded and executed by a processor to implement the above-mentioned infectious disease prediction method and / or the training method of the graph convolutional neural network.

[0036] According to another aspect of the present application, there is provided a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above-mentioned infectious disease prediction method and / or the training method of the graph convolutional neural network.

[0037] The beneficial effects brought by the technical solutions provided in the embodiments of the present application at least include:

[0038] By using the first physiological information and the first location information of the first user to update the basic graph network and obtain the first user feature vector, and converting the first user feature vector into the probability of the first user having an infectious disease through the probability calculation network, the above method not only uses the physiological information of the first user, but also uses the location information of the first user to predict the disease probability. Combining the spatial transmission of infectious diseases, the accuracy of infectious disease prediction is significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0040] Figure 1It is a schematic diagram of a computer system provided by an exemplary embodiment of the present application;

[0041] Figure 2 It is a flowchart of an infectious disease prediction method provided by an exemplary embodiment of the present application;

[0042] Figure 3 It is provided by an exemplary embodiment of the present application and applies the Figure 2 interface diagram of a product of the infectious disease prediction method shown;

[0043] Figure 4 It is a flowchart of an infectious disease prediction method provided by another exemplary embodiment of the present application;

[0044] Figure 5 It is a schematic diagram of a first graph network provided by an exemplary embodiment of the present application;

[0045] Figure 6 It is a flowchart of an infectious disease prediction method of another exemplary embodiment of the present application;

[0046] Figure 7 It is a schematic diagram of a second graph network provided by an exemplary embodiment of the present application;

[0047] Figure 8 It is a flowchart of an infectious disease prediction method of another exemplary embodiment of the present application;

[0048] Figure 9 It is a schematic diagram of a third graph network provided by an exemplary embodiment of the present application;

[0049] Figure 10 It is a flowchart of an infectious disease prediction method of another exemplary embodiment of the present application;

[0050] Figure 11 It is a schematic diagram of a fourth graph network provided by an exemplary embodiment of the present application;

[0051] Figure 12 It is a schematic diagram of a fifth graph network and a sixth graph network provided by an exemplary embodiment of the present application;

[0052] Figure 13 It is a flowchart of a training method of a graph convolutional neural network provided by an exemplary embodiment of the present application;

[0053] Figure 14 It is a flowchart of a training method of a graph convolutional neural network provided by another exemplary embodiment of the present application;

[0054] Figure 15 It is a structural block diagram of an infectious disease prediction device provided by an exemplary embodiment of the present application;

[0055] Figure 16 It is a structural block diagram of a training device for a graph convolutional neural network provided by an exemplary embodiment of the present application;

[0056] Figure 17 It is a structural block diagram of a training device for a graph convolutional neural network provided by another exemplary embodiment of the present application;

[0057] Figure 18 It is a structural block diagram of a computer device provided by an exemplary embodiment of the present application. Detailed implementation manners

[0058] To make the objectives, technical solutions, and advantages of the present application clearer, the following will further describe the embodiments of the present application in detail with reference to the accompanying drawings.

[0059] First, briefly introduce the nouns involved in the embodiments of the present application:

[0060] Graph network: In the embodiments of the present application, a graph network refers to data stored in the form of a graph. A graph network is also called graph data, a graph model, a graph representation, or graph-structured data. A graph network includes at least one node and at least one edge. Each node has a corresponding feature, and the edge is used to represent the connection relationship between different nodes.

[0061] Graph neural network: It refers to a general term for models in which a neural network is applied to a graph network. Graph neural networks include graph convolutional neural networks (Graph Convolutional Network, GCN), graph attention networks, etc. A graph neural network is used to predict the category of a graph based on the structural features of the graph. Specifically, a graph neural network will include one or more feature extraction layers. The feature extraction layer is, for example, a graph convolutional layer (GraphConvolution Layers, GCL). The feature extraction layer is used to extract the structural features of the graph. If two graphs are isomorphic, then the graphs of the two graphs will have similar structural features after passing through the feature extraction layer. If two graphs are heterogeneous, then the graphs of the two graphs will have different structural features after passing through the feature extraction layer. Therefore, a graph neural network can map graph structures with isomorphic properties to the same representation domain and output the same category.

[0062] Graph Convolutional Neural Network: A type of graph neural network that uses graph convolution. A graph convolutional neural network includes at least one graph convolutional layer. The role of the graph convolutional layer is similar to that of a feature extractor, and the object of feature extraction is the graph, and the extracted features are the structural features contained in the graph. Specifically, the graph convolutional layer includes multiple convolutional operators, also known as convolution kernels. Essentially, a convolution kernel can be a weight matrix, and the weight values in the weight matrix are obtained through the model training stage. Each weight matrix formed by the weight values obtained through training can be used to extract features from the input graph, so that the graph convolutional neural network can make correct predictions during the application stage.

[0063] The graph convolutional layer realizes the function of feature extraction through graph convolution processing. Graph convolution processing is an operation that performs a non-linear transformation on the input data. For the first graph convolutional layer of the graph convolutional neural network, the input data of the graph convolution processing is the graph; for the second convolutional layer to the last graph convolutional layer, the input data of the graph convolution processing is the output result of the previous graph convolutional layer.

[0064] The solution of the embodiment of the present application includes the training stage and the usage stage of the graph convolutional neural network. Figure 1 Figure 8 shows the computer systems of the training device 101 and the usage device 102 of the graph convolutional neural network provided by an exemplary embodiment of the present application. As Figure 1 shown, the graph convolutional neural network is trained by the training device 101 of the graph convolutional neural network, and the graph convolutional neural network is sent to the usage device 102, and the graph convolutional neural network can be used in the usage device 102 of the graph convolutional neural network.

[0065] Among them, the above-mentioned training device 101 and usage device 102 of the graph convolutional neural network can be computer devices with machine learning capabilities. For example, the computer device can be a terminal or a server.

[0066] Optionally, the training device 101 and the using device 102 of the above graph convolutional neural network may be the same computer device, or the training device 101 and the using device 102 of the graph convolutional neural network may also be different computer devices. Moreover, when the training device 101 of the graph convolutional neural network and the using device 102 of the graph convolutional neural network are different devices, the training device 101 of the graph convolutional neural network and the using device 102 of the graph convolutional neural network may be devices of the same type. For example, the training device 101 of the graph convolutional neural network and the using device 102 of the graph convolutional neural network may both be servers; or, the training device 101 of the graph convolutional neural network and the using device 102 of the graph convolutional neural network may also be devices of different types. The above server may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or may also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. The above terminal may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected through wired or wireless communication methods, and this application does not make any restrictions here.

[0067] To improve the accuracy of infectious disease prediction, Figure 2 is a flowchart of an infectious disease prediction method provided by an exemplary embodiment of the present application. In this embodiment, it is exemplified that the method is executed by Figure 1 the terminal (using device 102) shown, and the method includes:

[0068] Step 220, obtain the first physiological information and the first location information of the first user;

[0069] The first user: refers to the user for whom infectious disease prediction is performed. Optionally, the first user is a user whose infectious disease status is not yet determined, that is, the first user does not carry the label of the infectious disease (the label is used to determine whether the user has an infectious disease); optionally, the first user is a user whose infectious disease status has been determined. At this time, the first user performs infectious disease prediction again to verify the determined result.

[0070] The first physiological information: is the physiological information of the first user related to infectious diseases. Schematically, the physiological information related to infectious diseases includes at least one of gender, age, symptoms, laboratory test index information, and imaging examination information.

[0071] First location information: It indicates the location where the first user is located. In this application, there is no limitation on the minimum unit of the location indicated by the location information. For example, the first location information indicates that the location where the first user is located is Shenzhen City, or the first location information indicates that the location where the first user is located is Nanyuan Street, Shenzhen City, or the first location information indicates that the location where the first user is located is the house with the house number 5 on Nanyuan Street, Shenzhen City.

[0072] Step 240: Based on the first physiological information and the first location information, add a first user node to the basic graph network of the infectious disease to obtain an updated graph network.

[0073] Basic graph network: It refers to a reusable graph network related to infectious diseases. The basic graph network includes at least one user node and / or at least one area node. Among them, the user node identifies the user suspected of having an infectious disease, and the area node identifies the area where the user suspected of having an infectious disease is located.

[0074] Optionally, the area refers to an administrative area obtained based on administrative planning; optionally, the area refers to an area obtained based on a specific division. For example, during the occurrence period of an infectious disease, the area within 3 km around Village A is marked as a high-risk area, and the high-risk area is the area where the area node identifies the user suspected of having an infectious disease.

[0075] Specifically, the details of adding the first user node to the basic graph network of the infectious disease are described in the following four possible implementation manners.

[0076] Step 260: Input the updated graph network into a graph convolutional neural network to predict the first user feature vector of the first user.

[0077] The graph convolutional neural network is used to extract features from the updated graph network and predict the first user feature vector of the first user. The first user feature vector is a feature vector obtained based on the first physiological information and the first location information of the first user. Specifically, inputting the updated graph network into the graph convolutional neural network to predict the first user feature vector of the first user is described in the following four possible implementation manners.

[0078] Step 280: Based on the first user feature vector, predict the probability that the first user has an infectious disease through a probability calculation network.

[0079] In one embodiment, the first user feature vector is passed through a normalization function to obtain a value in the range of [0, 1], and this value is the probability that the first user has an infectious disease. Optionally, the normalization function is the sigmoid function.

[0080] In one embodiment, Figure 3 shows the application of Figure 2Product interface diagram of the infectious disease prediction method shown, where the risk score 301 shows the disease probabilities of the first patient, the second patient, the third patient, and the fourth patient.

[0081] In summary, the above method updates the basic graph network through the first physiological information and the first location information of the first user, and obtains the first user feature vector. Then, through the probability calculation network, the first user feature vector is converted into the probability that the first user has an infectious disease. The above method not only uses the physiological information of the first user, but also uses the location information of the first user to predict the disease probability. Combining the spatial transmission of infectious diseases, it significantly improves the accuracy of infectious disease prediction.

[0082] The first possible implementation: Based on Figure 2 In the optional embodiment shown, Figure 4 The flowchart of the infectious disease prediction method provided by an exemplary embodiment of the present application is shown. Among them, step 240 can be replaced by: step 242-1 and step 242-2, and step 260 can be replaced by step 262-1, step 262-2, step 262-3, and step 262-4.

[0083] Step 242-1, based on the first physiological information, generate a first user node in the basic graph network;

[0084] Among them, the basic graph network includes a first region node, and the first region node identifies the first region where the user suspected of having an infectious disease is located. Based on the first physiological information, the terminal generates a first user node in the basic graph network, and the first user node carries the first physiological information of the first user.

[0085] Step 242-2, when the location where the first user is located indicated by the first location information falls within the first region, perform an edge connection operation between the first user node and the first region node in the basic graph network to form a first graph network;

[0086] In an embodiment, the location where the first user is located indicated by the first location information is Nanyuan Street, Shenzhen City, and the first region is Shenzhen City. That is, the location where the first user is located indicated by the first location information falls within the first region. Perform an edge connection operation between the first user node and the first region node in the basic graph network to form a first graph network (i.e., the updated graph network).

[0087] Schematically, Figure 5 The first graph network is shown, where there is a connection between the first region node 501 and the first user node 502.

[0088] Step 262-1: Obtain an initial regional feature vector based on the historical disease information carried by the first regional node in the first graph network; obtain an initial user feature vector based on the first physiological information carried by the first user node in the first graph network.

[0089] Optionally, the historical disease information includes the historical confirmed number of infectious diseases within a certain past time period.

[0090] In one embodiment, the first regional node is represented by d, and the first user node is represented by u. Based on the historical disease information carried by the first regional node in the first graph network, the terminal obtains an initial regional feature vector. Based on the first physiological information carried by the first user node in the first graph network, obtain an initial user feature vector.

[0091] In one embodiment, the initial user feature vector includes at least one of a yes / no type feature (symptom) represented by 0 or 1, a numerical type feature (age) represented by a floating point number, and a categorical type feature (gender) represented by a one-hot vector.

[0092] Step 262-2: Calculate and obtain the first-layer regional feature vector based on the initial regional feature vector and the initial user feature vector; calculate and obtain the first-layer user feature vector based on the initial regional feature vector and the initial user feature vector.

[0093] In one embodiment, the update formula for the feature vector of the first regional node is:

[0094]

[0095] where φ represents an activation function. represents the information flow from the first regional node itself in the i-th layer. represents the information flow from the first user node u in the i-th layer.

[0096]

[0097]

[0098] where represents the i-1 layer regional feature vector. represents the i-1 layer user feature vector. represents the first parameter matrix of the first regional node in the i-th layer. represents the second parameter matrix of the first regional node in the i-th layer.

[0099] In one embodiment, the update formula for the feature vector of the first user node is:

[0100]

[0101] Among them, φ represents the activation function, represents the information flow from the first user node of the i-th layer itself, represents the information flow from the first regional node of the i-th layer.

[0102]

[0103]

[0104] Among them, represents the regional feature vector of the (i - 1)-th layer, represents the user feature vector of the (i - 1)-th layer, represents the third parameter matrix of the i-th layer of the first user node, represents the fourth parameter matrix of the i-th layer of the first user node.

[0105] When i = 1, the regional feature vector of the first layer is based on the initial regional feature vector and the initial user feature vector obtained, and the user feature vector of the first layer is based on the initial regional feature vector and the initial user feature vector obtained.

[0106] Step 262-3: Based on the regional feature vector of the (i - 1)-th layer and the user feature vector of the (i - 1)-th layer, calculate the regional feature vector of the i-th layer; based on the regional feature vector of the (i - 1)-th layer and the user feature vector of the (i - 1)-th layer, calculate the user feature vector of the i-th layer;

[0107] Combined with reference to the above formulas (1) to (6), the regional feature vector of the i-th layer is based on the regional feature vector of the (i - 1)-th layer and the user feature vector of the (i - 1)-th layer obtained, and the user feature vector of the i-th layer is based on the initial regional feature vector and the user feature vector of the i-th layer obtained.

[0108] Step 262-4: Output the user feature vector of the k-th layer as the first user feature vector.

[0109] In one embodiment, the graph convolutional neural network includes k graph convolutional layers, and the terminal outputs the user feature vector of the k-th layer output by the above k-th graph convolutional layer as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

[0110] In summary, the above method updates the basic graph network to obtain the updated first graph network. The first graph network includes the first regional node and the first user node, realizing the data integration of the physiological information and location information of the user through the graph network. Furthermore, the disease probability of the first user can be obtained through the graph convolutional neural network and the probability calculation network.

[0111] The second possible implementation manner: Based on Figure 2 In the optional embodiment shown, Figure 6 FIG. shows a flowchart of an infectious disease prediction method provided by an exemplary embodiment of the present application. Among them, step 240 can be replaced by: step 244-1 and step 244-2, and step 260 can be replaced by step 264-1, step 264-2, step 264-3 and step 264-4.

[0112] Step 244-1: Generate a first user node in the basic graph network based on the first physiological information;

[0113] Wherein, the basic graph network includes a first regional node and a second regional node connected to the first regional node. The first regional node identifies the first region where the user suspected of having an infectious disease is located, and the second regional node identifies the second region where the user suspected of having an infectious disease is located.

[0114] Moreover, the number of population flows between the second regional node and the first regional node reaches a quantity threshold, and / or the distance between the regions indicated by the second regional node and the first regional node is lower than the first distance threshold.

[0115] Illustratively, the first region is Shenzhen and the second region is Guangzhou. Based on the fact that the number of population flows between Shenzhen and Guangzhou reaches the quantity threshold and the distance between Shenzhen and Guangzhou is lower than the first distance threshold, the first regional node and the second regional node are connected in the basic graph network.

[0116] In one embodiment, the terminal generates a first user node in the basic graph network based on the first physiological information, and the first user node carries the first physiological information.

[0117] Step 244-2: In the case where the location of the first user indicated by the first location information falls within the first region, perform an edge connection operation between the first user node and the first regional node in the basic graph network to form a second graph network;

[0118] In one embodiment, the location of the first user of the first location information is Nanyuan Street, Shenzhen, and the first region is Shenzhen. That is, the location of the first user indicated by the first location information falls within the first region. An edge connection operation is performed between the first user node and the first regional node in the basic graph network to form a second graph network (i.e., the updated graph network).

[0119] Schematic Figure 7 Figure 7 shows a second graph network, in which the first regional node 701 is connected to the second regional node 702, and the first regional node 701 is connected to the first user node 703.

[0120] Step 264-1: Based on the historical disease information carried by the first regional node in the second graph network, obtain the initial regional feature vector of the first regional node; based on the historical disease information carried by the second regional node in the second graph network, obtain the initial regional feature vector of the second regional node; based on the first physiological information carried by the first user node in the second graph network, obtain the initial user feature vector;

[0121] In one embodiment, the first regional node is represented by d, the second regional node is represented by d’, and the first user node is represented by u. Based on the historical disease information carried by the first regional node in the second graph network, the terminal obtains the initial regional feature vector of the first regional node Based on the historical disease information carried by the second regional node in the second graph network, the terminal obtains the initial regional feature vector of the second regional node Based on the first physiological information carried by the first user node in the second graph network, obtain the initial user feature vector

[0122] Step 264-2: Based on the initial regional feature vector of the first regional node, the initial regional feature vector of the second regional node, and the initial user feature vector, combined with the first weight, calculate the first-layer regional feature vector of the first regional node; based on the initial regional feature vector of the first regional node and the initial user feature vector, calculate the first-layer user feature vector;

[0123] Among them, the first weight is associated with the population flow quantity between the first regional node and the second regional node.

[0124] In one embodiment, the update formula for the feature vector of the first regional node is:

[0125]

[0126] Among them, φ represents the activation function, represents the information flow from the first regional node d itself in the i-th layer, represents the information flow from the first user node u in the i-th layer, represents the information flow from the second regional node d′ in the i-th layer, α d←d′ represents the first weight, N d (d) represents the set of all regional nodes connected to the first regional node d.

[0127]

[0128]

[0129]

[0130]

[0131] Among them, d″ represents the regional nodes (including the second regional node) connected to the first regional node d, and w d←d′ represents the edge weight between the second regional node d′ and the first regional node d (obtained based on the population flow quantity between the first region and the second region). represents the fifth parameter matrix of the i-th layer of the first regional node, represents the sixth parameter matrix of the i-th layer of the first regional node, represents the seventh parameter matrix of the i-th layer of the first regional node.

[0132] In one embodiment, the update formula for the feature vector of the first user node is:

[0133]

[0134] Among them, φ represents the activation function, represents the information flow from the first user node itself in the i-th layer, represents the information flow from the first regional node in the i-th layer.

[0135]

[0136]

[0137] Among them, represents the regional feature vector of the (i - 1)-th layer of the first regional node, represents the user feature vector of the (i - 1)-th layer, represents the eighth parameter matrix of the i-th layer of the first user node, represents the ninth parameter matrix of the i-th layer of the first user node.

[0138] When i = 1, the regional feature vector of the first layer of the first regional node is obtained based on the initial regional feature vector of the first regional node the initial regional feature vector of the second regional node and the initial user feature vector The user feature vector of the first layer is obtained based on the initial regional feature vector and the initial user feature vector Obtained.

[0139] Step 264-3: Based on the (i-1)-th layer regional feature vector of the first regional node, the (i-1)-th layer regional feature vector of the second regional node, and the (i-1)-th layer user feature vector, combined with the first weight, calculate to obtain the i-th layer regional feature vector of the first regional node; based on the (i-1)-th layer regional feature vector of the first regional node and the (i-1)-th layer user feature vector, calculate to obtain the i-th layer user feature vector.

[0140] Combined with reference to the above formulas (7) to (14), the i-th layer regional feature vector of the first regional node is based on the (i-1)-th layer regional feature vector of the first regional node the (i-1)-th layer regional feature vector of the second regional node and the (i-1)-th layer user feature vector obtained by combining the first weight. The i-th layer user feature vector is based on the (i-1)-th layer regional feature vector of the first regional node and the (i-1)-th layer user feature vector Obtained.

[0141] Step 264-4: Output the k-th layer user feature vector as the first user feature vector.

[0142] In one embodiment, the graph convolutional neural network includes k graph convolutional layers, and the terminal outputs the k-th layer user feature vector output by the above k-th graph convolutional layer as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

[0143] In summary, the above method updates the basic graph network to obtain an updated second graph network, which includes a first regional node, a second regional node, and a first user node, realizing data integration of the user's physiological information and location information through the graph network, and further obtaining the disease probability of the first user through the graph convolutional neural network and the probability calculation network.

[0144] The third possible implementation manner: Based on Figure 2 in the optional embodiment shown, Figure 8 shows a flowchart of an infectious disease prediction method provided by an exemplary embodiment of the present application. Among them, step 240 can be replaced by: step 246-1 and step 246-2, and step 260 can be replaced by step 266-1, step 266-2, step 266-3, and step 266-4.

[0145] Step 246-1: Based on the first physiological information, generate a first user node in the basic graph network.

[0146] The basic graph network includes a first regional node and a second user node connected to the first regional node. The first regional node identifies the first region where the user suspected of having an infectious disease is located, and the second user node identifies the second user suspected of having an infectious disease.

[0147] The second user node: refers to the user node that already existed before the update of the basic graph network.

[0148] In one embodiment, the terminal generates a first user node in the basic graph network based on the first physiological information, and the first user node carries the first physiological information.

[0149] Step 246-2, when the location where the first user indicated by the first location information is located falls within the first region, perform an edge connection operation between the first user node and the first regional node in the basic graph network to form a third graph network;

[0150] In one embodiment, the location where the first user of the first location information is located is Nanyuan Street, Shenzhen City, and the first region is Shenzhen City. That is, the location where the first user indicated by the first location information falls within the first region. Perform an edge connection operation between the first user node and the first regional node in the basic graph network to form a third graph network (i.e., the updated graph network).

[0151] Schematically, Figure 9 The third graph network is shown, where there is a connection between the first regional node 901 and the first user node 902, and a connection between the first regional node 901 and the second user node 903. Among them, the location where the second user indicated by the second user node falls within the first region.

[0152] Step 266-1, based on the historical disease information carried by the first regional node in the third graph network, obtain an initial regional feature vector; based on the first physiological information carried by the first user node in the third graph network, obtain an initial user feature vector of the first user node; based on the second physiological information carried by the second user node in the third graph network, obtain an initial user feature vector of the second user node;

[0153] In one embodiment, the first regional node is represented by d, and a user node in the user node set jointly constituted by the first user node and at least one second user node is represented by u i represented, the first user node is represented by u, and based on the historical disease information carried by the first regional node in the third graph network, the terminal obtains the initial regional feature vector of the first regional node Based on the first physiological information carried by the first user node in the third graph network, obtain the initial user feature vector

[0154] Step 266-2: Calculate the first-layer regional feature vector based on the initial regional feature vector, the initial user feature vector of the first user node, and the initial user feature vector of the second user node; calculate the first-layer user feature vector of the first user node based on the initial regional feature vector and the initial user feature vector of the first user node.

[0155] In one embodiment, the update formula for the feature vector of the first regional node is:

[0156]

[0157] where φ represents the activation function, represents the information flow from the first regional node d itself at the i-th layer, represents an information flow from a user node u in the set of user nodes connected to the first regional node d at the i-th layer (including the first user node and the second user node), i of, N u (d) represents the set of all user nodes connected to the first regional node d.

[0158]

[0159]

[0160] where, represents the (i - 1)-th layer regional feature vector, represents the user feature vector of a user node in the set of user nodes at the (i - 1)-th layer, represents the tenth parameter matrix of the first regional node at the i-th layer, represents the eleventh parameter matrix of the first regional node at the i-th layer.

[0161] In one embodiment, the update formula for the feature vector of the first user node is:

[0162]

[0163] where φ represents the activation function, represents the information flow from the first user node itself at the i-th layer, represents the information flow from the first regional node at the i-th layer.

[0164]

[0165]

[0166] where, represents the (i - 1)-th layer regional feature vector, represents the (i - 1)-th layer user feature vector, The twelfth parameter matrix of the i-th layer of the first user node, The thirteenth parameter matrix of the i-th layer of the first user node.

[0167] When i = 1, the regional feature vector of the first layer of the first regional node is based on the initial regional feature vector of the first regional node the initial user feature vector of the first user node and the initial user feature vector of the second user node, and the user feature vector of the first layer of the first user node is based on the initial regional feature vector and the initial user feature vector obtained.

[0168] Step 266-3, based on the regional feature vector of the (i-1)-th layer, the user feature vector of the (i-1)-th layer of the first user node, and the user feature vector of the (i-1)-th layer of the second user node, calculate to obtain the regional feature vector of the i-th layer; based on the regional feature vector of the (i-1)-th layer and the user feature vector of the (i-1)-th layer of the first user node, calculate to obtain the user feature vector of the i-th layer of the first user node;

[0169] [[ID=2-]]Combined with the above formulas (15) to (20), the regional feature vector of the i-th layer is based on the regional feature vector of the (i-1)-th layer of the first regional node the user feature vectors of the user node set composed of the first user node and the second user node of the (i-1)-th layer. The user feature vector of the i-th layer is based on the regional feature vector of the (i-1)-th layer of the first regional node and the user feature vector of the (i-1)-th layer obtained.

[0170] Step 266-4, output the user feature vector of the k-th layer of the first user node as the first user feature vector.

[0171] In one embodiment, the graph convolutional neural network includes k graph convolutional layers, and the terminal outputs the user feature vector of the k-th layer of the first user node output by the above k-th graph convolutional layer as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

[0172] In summary, the above method updates the basic graph network to obtain the updated third graph network, and the third graph network includes the first user node and the second user node, realizing the data integration of the physiological information and location information of the user through the graph network, and further obtaining the disease probability of the first user through the graph convolutional neural network and the probability calculation network.

[0173] The fourth possible implementation manner, based on Figure 2In the optional embodiment shown, Figure 10 FIG. 2 shows a flowchart of an infectious disease prediction method provided by an exemplary embodiment of the present application. Among them, step 240 can be replaced with: step 248-1 and step 248-2, and step 260 can be replaced with step 268-1, step 268-2, step 268-3, and step 268-4.

[0174] Step 248-1: Generate a first user node in the basic graph network based on the first physiological information.

[0175] The basic graph network includes a third user node, which identifies a third user suspected of having an infectious disease. The third user node carries the third physiological information and the third location information of the third user.

[0176] Step 248-2: When the similarity between the first physiological information and the third physiological information is greater than the similarity threshold, and / or the distance between the positions indicated by the first location information and the third location information is less than the second distance threshold, perform an edge connection operation between the first user node and the third user node in the basic graph network to form a fourth graph network.

[0177] In one embodiment, the first physiological information indicates the symptoms of the first user, and the third physiological information indicates the symptoms of the third user. When the similarity between the symptoms of the first user and the third user is greater than the similarity threshold, and the distance between the location where the first user is located and the location where the third user is located is less than the second distance threshold, perform an edge connection operation between the first user node and the third user node in the basic graph network to form a fourth graph network. Schematically, Figure 11 FIG. 4 shows the fourth graph network, where the first user node 1101 and the third user node 1102 are connected.

[0178] Step 268-1: Obtain the initial user feature vector of the first user node based on the first physiological information carried by the first user node in the fourth graph network; obtain the initial user feature vector of the third user node based on the third physiological information carried by the third user node in the fourth graph network.

[0179] In one embodiment, the first user node is represented by m, and the second user node is represented by n. Obtain the initial user feature vector of the first user node m based on the first physiological information carried by the first user node. Obtain the initial user feature vector of the third user node n based on the third physiological information carried by the third user node.

[0180] Step 268-2: Based on the initial user feature vectors of the first user node and the third user node, and combined with the second weight, calculate the first-layer user feature vectors of the first user node and the third user node;

[0181] Among them, the second weight is associated with at least one of the similarity between the first physiological information and the third physiological information and the distance between the positions indicated by the first location information and the third location information;

[0182] In one embodiment, the update formula for the feature vector of the first user node is:

[0183]

[0184] Among them, σ represents the activation function, and α mn is the second weight, W 14 represents the fourteenth weight of the first user node in the i-th layer, and W 15 represents the fifteenth weight of the third user node in the i-th layer.

[0185]

[0186]

[0187] Among them, a T represents a specific parameter matrix, which is trained together with W 16 during the training process. N(m) represents the set of user nodes connected to the first user node, represents the normalized distance, δ d represents the second distance threshold, represents the distance between the locations where the first user and the third user are located, s mn represents the similarity between the first physiological information of the first user and the third physiological information of the third user. [] represents vector concatenation, and W16 represents the sixteenth parameter matrix of the first user node in the i-th layer.

[0188]

[0189]

[0190] w f represents the weight of a physiological information f (one of the physiological information in the first physiological information), S m represents the first physiological information of the first user, S n represents the third physiological information of the third user, N f represents the number of confirmed patients with the physiological information f, and N represents the number of currently confirmed patients with infectious diseases.

[0191] Similarly, referring to the above formula (21), the eigenvector of the second user node n can be calculated.

[0192] Step 268-3: Based on the user feature vector of the i-1th layer of the first user node and the user feature vector of the i-1th layer of the third user node, and combining with the second weight, calculate the user feature vector of the i-th layer of the first user node and the user feature vector of the i-th layer of the third user node.

[0193] Combining with reference to the above formulas (21), (22), and (23), the user feature vector of the i-th layer of the first user node and the user feature vector of the i-th layer of the third user node can be calculated.

[0194] Step 268-4: Output the user feature vector of the k-th layer of the first user node as the first user feature vector.

[0195] In one embodiment, the graph convolutional neural network includes k graph convolutional layers. The terminal outputs the user feature vector of the k-th layer of the first user node output by the above k-th graph convolutional layer as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

[0196] In summary, the above method updates the basic graph network to obtain an updated fourth graph network, which includes a first user node and a third user node, realizing the data integration of the physiological information and location information of the user through the graph network. Further, the disease probability of the first user can be obtained through the graph convolutional neural network and the probability calculation network.

[0197] It should be noted that the above first possible implementation manner to the fourth possible implementation manner can be combined. Schematically, Figure 12 shows a schematic diagram of a fifth graph network and a sixth graph network provided by an exemplary embodiment of the present application.

[0198] The fifth graph network (a) is obtained by combining the first graph network, the second graph network, the third graph network, and the fourth graph network. The first region node 1201 is connected to the first user node 1203, the first region node 1201 is connected to the second region node 1202, the first region node 1201 is further connected to at least one second user node, and the first user node 1203 is further connected to at least one third user node.

[0199] The sixth graph network (b) is obtained by combining the first graph network, the second graph network, and the third graph network. The first region node 1201 is connected to the first user node 1203, the first region node 1201 is connected to the second region node 1202, and the first region node 1201 is further connected to at least one second user node.

[0200] Another point worth noting is that all the above parameter matrices are updated through model training.

[0201] In one embodiment, based on Figure 2 the optional embodiment shown, step 280 may be replaced with:

[0202] S1: Concatenate the first user feature vector and the initial user feature vector of the first user node to obtain a first intermediate vector;

[0203] S2: Input the first intermediate vector into a fully connected network to obtain a second intermediate vector;

[0204] S3: Input the second intermediate vector into an activation function to predict the probability that the first user has an infectious disease.

[0205] Schematically, the formula for calculating the probability that the first user has an infectious disease is as follows:

[0206]

[0207] where FC represents the fully connected network, σ represents the sigmoid function, represents the user feature vector output by the k-th layer of the first user node, represents the initial feature vector of the first user node, and [] represents vector concatenation.

[0208] In one embodiment, based on Figure 2 the optional embodiment shown, after step 280, it further includes:

[0209] Delete the first user node in the updated graph network. The first user node is a user node generated in the basic graph network based on the first physiological information of the first user.

[0210] To train the above graph convolutional neural network, Figure 13 shows a flowchart of a training method for a graph convolutional neural network provided by an exemplary embodiment of the present application. Taking the method applied to Figure 1 the terminal (training device 101) of the graph convolutional neural network shown as an example, this training method adopts a semi-supervised mode, and the method includes:

[0211] Step 1301, obtain the sample physiological information and sample location information of n sample users;

[0212] where the n sample users include n1 first sample users and n2 second sample users. The first sample users carry labels, and the labels are used to determine whether the first sample users have infectious diseases. The second sample users do not carry labels.

[0213] In one embodiment, the terminal obtains the sample physiological information and sample location information of n sample users.

[0214] Step 1302: Input the sample graph network corresponding to the sample physiological information and sample location information of n sample users into a graph convolutional neural network, and calculate n sample feature vectors of n sample users through the graph convolutional neural network;

[0215] In one embodiment, the terminal inputs the sample graph network corresponding to the sample physiological information and sample location information of n sample users into a graph convolutional neural network, and calculates n sample feature vectors of n sample users through the graph convolutional neural network.

[0216] Optionally, the terminal generates n sample nodes based on the sample physiological information of n sample users. Based on the positions indicated by the n sample location information falling within the first area indicated by the first area node, the terminal connects the n sample nodes to the first area node to obtain a first sample graph network;

[0217] Optionally, the terminal generates n sample nodes based on the sample physiological information of n sample users. Based on the positions indicated by the n sample location information falling within the first area indicated by the first area node or the second area indicated by the second area node, the terminal connects the n sample nodes to the first area node or the second area node to obtain a second sample graph network;

[0218] Optionally, the terminal generates n sample nodes based on the sample physiological information of n sample users. Based on the positions indicated by the n sample location information falling within the first area indicated by the first area node, the terminal connects the n sample nodes to the first area node and jointly forms a third sample graph network with the second user node. The position where the second user indicated by the second user node is located falls within the first area indicated by the first area node;

[0219] Optionally, based on the distance between one sample user among the n sample users and the third user indicated by the third user node being lower than a second distance threshold, and / or, based on the similarity between the physiological information of one sample user among the n sample users and the third physiological information of the third user indicated by the third user node reaching a similarity threshold, the terminal connects the n sample nodes to the third user node to obtain a fourth sample graph network.

[0220] Combined with reference Figure 5 、 Figure 7 、 Figure 9 、 Figure 11 and Figure 12, the first sample graph network, the second sample graph network, the third sample graph network, and the fourth sample graph network are respectively similar to the first graph network, the second graph network, the third graph network, and the fourth graph network;

[0221] It should be noted that based on the above-mentioned first sample graph network, second sample graph network, third sample graph network, and fourth sample graph network, a new sample graph network can also be constructed, and the new sample graph network can be similar to the fifth graph network and the sixth graph network.

[0222] Step 1303, for one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has an infectious disease through the probability calculation network;

[0223] In one embodiment, for one of the n1 first sample users, the terminal obtains the predicted probability that the first sample user has an infectious disease through the probability calculation network based on the sample feature vector of the first sample user.

[0224] In one embodiment, the terminal splices the sample feature vector of the first sample user and the initial feature vector of the first sample user to obtain a first sample intermediate vector; then, the terminal inputs the first sample intermediate vector into a fully connected network to obtain a second sample intermediate vector; finally, the terminal inputs the second sample intermediate vector into an activation function to obtain the probability that the first sample user has an infectious disease.

[0225] Step 1304, train the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0226] In one embodiment, the terminal trains the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0227] In summary, the above training method trains the graph convolutional neural network by setting that n1 of the n sample users are first sample users with labels and n2 of the second sample users have no labels, so that the sample users without labels still participate in the integrated calculation of the data, but the predicted probabilities of the sample users without labels do not participate in the training of the graph convolutional neural network.

[0228] To train the above graph convolutional neural network, Figure 14 shows a flowchart of a training method for a graph convolutional neural network provided by an exemplary embodiment of the present application, taking this method applied to Figure 1 the terminal (training device 101) of the graph convolutional neural network shown as an example. This training method adopts a supervised mode, and this method includes:

[0229] Step 1401: Obtain the sample physiological information and sample location information of n1 first-sample users;

[0230] Among them, the first-sample users carry tags, and the tags are used to determine whether the first-sample users have infectious diseases; in one embodiment, the terminal obtains the sample physiological information and sample location information of n1 first-sample users

[0231] Step 1402: Input the sample graph network corresponding to the sample physiological information and sample location information of n1 first-sample users into a graph convolutional neural network, and calculate n1 sample feature vectors of n1 first-sample users through the graph convolutional neural network;

[0232] In one embodiment, the terminal inputs the sample physiological information and sample location information of n1 first-sample users into a graph convolutional neural network, and calculates n1 sample feature vectors of n1 first-sample users through the graph convolutional neural network.

[0233] Optionally, the terminal generates n sample nodes based on the sample physiological information of n1 first-sample users. Based on the positions indicated by the n sample location information falling within the first area indicated by the first area node, the terminal connects the n sample nodes with the first area node to obtain a fifth sample graph network;

[0234] Optionally, the terminal generates n sample nodes based on the sample physiological information of n1 first-sample users. Based on the positions indicated by the n sample location information falling within the first area indicated by the first area node or the second area indicated by the second area node, the terminal connects the n sample nodes with the first area node or the second area node to obtain a sixth sample graph network;

[0235] Optionally, the terminal generates n sample nodes based on the sample physiological information of n1 first-sample users. Based on the positions indicated by the n sample location information falling within the first area indicated by the first area node, the terminal connects the n sample nodes with the first area node and jointly forms a seventh sample graph network with the second user node, and the position where the second user indicated by the second user node is located falls within the first area indicated by the first area node;

[0236] Optionally, the terminal generates n sample nodes based on the sample physiological information of n1 first-sample users. Based on the distance between one of the n1 first-sample users and the third user indicated by the third user node being lower than the second distance threshold, and / or the similarity between the physiological information of one of the n1 first-sample users and the third physiological information of the third user indicated by the third user node reaching the similarity threshold, the terminal connects the n sample nodes with the third user node to obtain an eighth sample graph network.

[0237] Combined reference Figure 5 、 Figure 7 、 Figure 9 、 Figure 11 and Figure 12 , the fifth sample graph network, the sixth sample graph network, the seventh sample graph network, and the eighth sample graph network are respectively similar to the first graph network, the second graph network, the third graph network, and the fourth graph network;

[0238] It should be noted that based on the above-mentioned fifth sample graph network, sixth sample graph network, seventh sample graph network, and eighth sample graph network, a new sample graph network can also be constructed, and the new sample graph network can be similar to the fifth graph network and the sixth graph network.

[0239] Step 1403, for one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has an infectious disease through the probability calculation network;

[0240] In one embodiment, the terminal, for one of the n1 first sample users, based on the sample feature vector of the first sample user, obtains the predicted probability that the first sample user has an infectious disease through the probability calculation network.

[0241] In one embodiment, the terminal splices the sample feature vector of the first sample user and the initial feature vector of the first sample user to obtain a first sample intermediate vector; then, the terminal inputs the first sample intermediate vector into a fully connected network to obtain a second sample intermediate vector; finally, the terminal inputs the second sample intermediate vector into an activation function to obtain the probability that the first sample user has an infectious disease.

[0242] Step 1404, based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users, train the graph convolutional neural network.

[0243] In one embodiment, the terminal trains the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0244] In summary, the above training method uses n1 first sample users with labels to train the graph convolutional neural network, realizing the training of the graph convolutional neural network.

[0245] In one embodiment, Figure 15 shows a structural block diagram of an infectious disease prediction device provided by an exemplary embodiment of the present application. The device includes:

[0246] An acquisition module 1501, configured to acquire the first physiological information and the first location information of the first user. The first physiological information is the physiological information related to the infectious disease of the first user, and the first location information indicates the location where the first user is located;

[0247] An update module 1502, configured to add a first user node to a basic graph network of infectious diseases based on first physiological information and first location information, to obtain an updated graph network; the basic graph network includes at least one user node and / or at least one area node, the user node identifies a user suspected of having an infectious disease, and the area node identifies the area where the user suspected of having an infectious disease is located;

[0248] A prediction module 1503, configured to input the updated graph network into a graph convolutional neural network, and predict to obtain a first user feature vector of the first user; and, based on the first user feature vector, predict the probability that the first user has an infectious disease through a probability calculation network.

[0249] In an optional embodiment, the basic graph network includes a first area node, and the updated graph network includes a first graph network, and the first area node identifies the first area where the user suspected of having an infectious disease is located.

[0250] The update module 1502 is further configured to generate a first user node in the basic graph network based on the first physiological information.

[0251] The update module 1502 is further configured to, when the location where the first user is located indicated by the first location information falls within the first area, perform an edge connection operation between the first user node and the first area node in the basic graph network to form a first graph network.

[0252] In an optional embodiment, the graph convolutional neural network has k graph convolutional layers.

[0253] In an optional embodiment, the prediction module 1503 is further configured to obtain an initial area feature vector based on historical disease information carried by the first area node in the first graph network; and obtain an initial user feature vector based on the first physiological information carried by the first user node in the first graph network.

[0254] In an optional embodiment, the prediction module 1503 is further configured to calculate a first-layer area feature vector based on the initial area feature vector and the initial user feature vector; and calculate a first-layer user feature vector based on the initial area feature vector and the initial user feature vector.

[0255] In an optional embodiment, the prediction module 1503 is further configured to calculate an i-th layer area feature vector based on the (i - 1)-th layer area feature vector and the (i - 1)-th layer user feature vector; and calculate an i-th layer user feature vector based on the (i - 1)-th layer area feature vector and the (i - 1)-th layer user feature vector.

[0256] In an optional embodiment, the prediction module 1503 is further configured to output the k-th layer user feature vector as the first user feature vector.

[0257] In an alternative embodiment, the basic graph network includes a first regional node and a second regional node connected to the first regional node, and the updated graph network includes a second graph network. The first regional node identifies a first region where a user suspected of having an infectious disease is located, and the second regional node identifies a second region where the user suspected of having an infectious disease is located.

[0258] In an alternative embodiment, the update module 1502 is further configured to generate a first user node in the basic graph network based on the first physiological information.

[0259] In an alternative embodiment, when the location of the first user indicated by the first location information falls within the first region, the update module 1502 is further configured to perform an edge connection operation between the first user node and the first regional node in the basic graph network to form a second graph network.

[0260] Wherein, the number of population flows between the second regional node and the first regional node reaches a quantity threshold, and / or the distance between the regions indicated by the second regional node and the first regional node is lower than a first distance threshold.

[0261] In an alternative embodiment, the graph convolutional neural network includes k graph convolutional layers.

[0262] In an alternative embodiment, the prediction module 1503 is further configured to obtain an initial regional feature vector of the first regional node based on the historical disease information carried by the first regional node in the second graph network; obtain an initial regional feature vector of the second regional node based on the historical disease information carried by the second regional node in the second graph network; and obtain an initial user feature vector based on the first physiological information carried by the first user node in the second graph network.

[0263] In an alternative embodiment, the prediction module 1503 is further configured to calculate a first-layer regional feature vector of the first regional node based on the initial regional feature vector of the first regional node, the initial regional feature vector of the second regional node, and the initial user feature vector, in combination with a first weight, where the first weight is associated with the number of population flows between the first regional node and the second regional node; and calculate a first-layer user feature vector based on the initial regional feature vector of the first regional node and the initial user feature vector.

[0264] In an alternative embodiment, the prediction module 1503 is further configured to calculate the i-th layer regional feature vector of the first regional node based on the (i-1)-th layer regional feature vector of the first regional node, the (i-1)-th layer regional feature vector of the second regional node, and the (i-1)-th layer user feature vector, in combination with the first weight; and calculate the i-th layer user feature vector based on the (i-1)-th layer regional feature vector of the first regional node and the (i-1)-th layer user feature vector.

[0265] In an alternative embodiment, the prediction module 1503 is further configured to output the k-th layer user feature vector as the first user feature vector.

[0266] In an alternative embodiment, the basic graph network includes a first regional node and a second user node connected to the first regional node. The updated graph network includes a third graph network. The first regional node identifies the first region where the user suspected of having an infectious disease is located, and the second user node identifies the second user suspected of having an infectious disease.

[0267] In an alternative embodiment, the update module 1502 is further configured to generate a first user node in the basic graph network based on the first physiological information.

[0268] In an alternative embodiment, the update module 1502 is further configured to, when the location where the first user is located indicated by the first location information falls within the first region, perform an edge connection operation between the first user node and the first regional node in the basic graph network to form a third graph network.

[0269] Wherein, the location where the second user indicated by the second user node falls within the first region.

[0270] In an alternative embodiment, the graph convolutional neural network has k graph convolutional layers.

[0271] In an alternative embodiment, the prediction module 1503 is further configured to obtain an initial regional feature vector based on the historical disease information carried by the first regional node in the third graph network; obtain an initial user feature vector of the first user node based on the first physiological information carried by the first user node in the third graph network; and obtain an initial user feature vector of the second user node based on the second physiological information carried by the second user node in the third graph network.

[0272] In an alternative embodiment, the prediction module 1503 is further configured to calculate the first layer regional feature vector based on the initial regional feature vector, the initial user feature vector of the first user node, and the initial user feature vector of the second user node; and calculate the first layer user feature vector of the first user node based on the initial regional feature vector and the initial user feature vector of the first user node.

[0273] In an alternative embodiment, the prediction module 1503 is further configured to calculate the i-th layer regional feature vector based on the (i - 1)-th layer regional feature vector, the (i - 1)-th layer user feature vector of the first user node, and the (i - 1)-th layer user feature vector of the second user node; and calculate the i-th layer user feature vector of the first user node based on the (i - 1)-th layer regional feature vector and the (i - 1)-th layer user feature vector of the first user node.

[0274] In an alternative embodiment, the prediction module 1503 is further configured to output the k-th layer user feature vector of the first user node as the first user feature vector.

[0275] In an alternative embodiment, the basic graph network includes a third user node, and the updated graph network includes a fourth graph network. The third user node identifies a third user suspected of having an infectious disease, and the third user node carries the third physiological information and the third location information of the third user.

[0276] In an alternative embodiment, the update module 1502 is further configured to generate a first user node in the basic graph network based on the first physiological information.

[0277] In an alternative embodiment, the update module 1502 is further configured to perform an edge connection operation between the first user node and the third user node in the basic graph network to form a fourth graph network when the similarity between the first physiological information and the third physiological information is greater than the similarity threshold, and / or the distance between the positions indicated by the first location information and the third location information is less than the second distance threshold.

[0278] In an alternative embodiment, the graph convolutional neural network includes k layers of graph convolutional layers.

[0279] In an alternative embodiment, the prediction module 1503 is further configured to obtain the initial user feature vector of the first user node based on the first physiological information carried by the first user node in the fourth graph network; and obtain the initial user feature vector of the third user node based on the third physiological information carried by the third user node in the fourth graph network.

[0280] In an alternative embodiment, the prediction module 1503 is further configured to calculate the first layer user feature vector of the first user node and the first layer user feature vector of the third user node based on the initial user feature vector of the first user node and the initial user feature vector of the third user node, in combination with a second weight, where the second weight is associated with at least one of the similarity between the first physiological information and the third physiological information and the distance between the positions indicated by the first location information and the third location information.

[0281] In an alternative embodiment, the prediction module 1503 is further configured to calculate the user feature vector of the i-th layer of the first user node and the user feature vector of the i-th layer of the third user node based on the user feature vector of the (i - 1)-th layer of the first user node and the user feature vector of the (i - 1)-th layer of the third user node, in combination with the second weight.

[0282] In an alternative embodiment, the prediction module 1503 is further configured to output the user feature vector of the k-th layer of the first user node as the first user feature vector.

[0283] In an alternative embodiment, the apparatus further includes a deletion module 1504, configured to delete the first user node in the updated graph network.

[0284] In an alternative embodiment, the probability calculation network includes an activation function.

[0285] In an alternative embodiment, the prediction module 1503 is further configured to splice the first user feature vector and the initial user feature vector of the first user node to obtain a first intermediate vector.

[0286] In an alternative embodiment, the prediction module 1503 is further configured to input the first intermediate vector into a fully connected network to obtain a second intermediate vector.

[0287] In an alternative embodiment, the prediction module 1503 is further configured to input the second intermediate vector into an activation function to predict the probability that the first user has an infectious disease.

[0288] In summary, the above apparatus updates the basic graph network through the first physiological information and the first location information of the first user, and obtains the first user feature vector, and converts the first user feature vector into the probability that the first user has an infectious disease through the probability calculation network. The above apparatus not only uses the physiological information of the first user, but also uses the location information of the first user to predict the disease probability, and combines the spatial transmission of infectious diseases, significantly improving the accuracy of infectious disease prediction.

[0289] Figure 16 The block diagram of the training apparatus of the graph convolutional neural network provided by an exemplary embodiment of the present application is shown. The apparatus trains the graph convolutional neural network in a semi-supervised manner. The apparatus includes:

[0290] An acquisition module 1601, configured to acquire the sample physiological information and sample location information of n sample users; the n sample users include n1 first sample users and n2 second sample users. The first sample users carry labels, and the labels are used to determine whether the first sample users have infectious diseases. The second sample users do not carry labels;

[0291] The calculation module 1602 is configured to input the sample graph network corresponding to the sample physiological information and sample location information of n sample users into a graph convolutional neural network, and calculate n sample feature vectors of the n sample users through the graph convolutional neural network;

[0292] The calculation module 1602 is further configured to, for one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has an infectious disease through a probability calculation network;

[0293] The training module 1603 is configured to train the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0294] In an alternative embodiment, the probability calculation network includes an activation function.

[0295] In an alternative embodiment, the calculation module 1602 is further configured to splice the sample feature vector of the first sample user and the initial feature vector of the first sample user to obtain a first sample intermediate vector.

[0296] In an alternative embodiment, the calculation module 1602 is further configured to input the first sample intermediate vector into a fully connected network to obtain a second sample intermediate vector.

[0297] In an alternative embodiment, the calculation module 1602 is further configured to input the second sample intermediate vector into an activation function to obtain the probability that the first sample user has an infectious disease.

[0298] In summary, the above training device trains the graph convolutional neural network by setting that n1 of the n sample users are first sample users with labels and n2 of the n sample users are second sample users without labels, enabling the sample users without labels to still participate in the integrated calculation of data, but the predicted probabilities of the sample users without labels do not participate in the training of the graph convolutional neural network.

[0299] Figure 17 The block diagram of the structure of the training device of the graph convolutional neural network provided by an exemplary embodiment of the present application is shown. This device trains the graph convolutional neural network in a supervised manner, and this device includes:

[0300] The acquisition module 1701 is configured to acquire the sample physiological information and sample location information of n1 first sample users; the first sample users carry labels, and the labels are used to determine whether the first sample users have an infectious disease;

[0301] A calculation module 1702 is configured to input a sample graph network corresponding to the sample physiological information and sample location information of n1 first sample users into a graph convolutional neural network, and calculate n1 sample feature vectors of the n1 first sample users through the graph convolutional neural network;

[0302] The calculation module 1702 is further configured to, for one of the n1 first sample users, obtain a predicted probability that the first sample user has an infectious disease through a probability calculation network based on the sample feature vector of the first sample user;

[0303] A training module 1703 is configured to train the graph convolutional neural network based on a loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

[0304] In an optional embodiment, the probability calculation network includes an activation function.

[0305] In an optional embodiment, the calculation module 1702 is further configured to splice the sample feature vector of the first sample user and the initial feature vector of the first sample user to obtain a first sample intermediate vector.

[0306] In an optional embodiment, the calculation module 1702 is further configured to input the first sample intermediate vector into a fully connected network to obtain a second sample intermediate vector.

[0307] In an optional embodiment, the calculation module 1702 is further configured to input the second sample intermediate vector into an activation function to obtain the probability that the first sample user has an infectious disease.

[0308] In summary, the above training device uses n1 first sample users with labels to train the graph convolutional neural network, realizing the training of the graph convolutional neural network.

[0309] Figure 18 The structural block diagram of a computer device 1800 provided by an exemplary embodiment of the present application is shown. The computer device can be a terminal or a server. In this embodiment, it can be simply described as the terminal independently training the graph convolutional neural network and / or the terminal independently using the graph convolutional neural network, or, the server independently training the graph convolutional neural network and / or the server independently using the graph convolutional neural network, or, the terminal and the server jointly training an image retrieval model and / or the terminal and the server jointly using the graph convolutional neural network.

[0310] Generally, the computer device 1800 includes a processor 1801 and a memory 1802.

[0311] The processor 1801 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. The processor 1801 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 1801 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 1801 may be integrated with a GPU (Graphics Processing Unit), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 1801 may further include an AI processor, which is used to process computational operations related to machine learning.

[0312] The memory 1802 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 1802 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 1802 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 1801 to implement the prediction method for infectious diseases or the training method for the graph convolutional neural network provided in the method embodiments of the present application.

[0313] In some embodiments, the computer device 1800 may further optionally include: a peripheral device interface 1803 and at least one peripheral device. The processor 1801, the memory 1802, and the peripheral device interface 1803 may be connected through a bus or signal lines. Each peripheral device may be connected to the peripheral device interface 1803 through a bus, signal lines, or a circuit board. Specifically, the peripheral devices include at least one of a radio frequency circuit 1804, a display screen 1805, a camera assembly 1806, an audio circuit 1807, a positioning component 1808, and a power supply 1809.

[0314] The peripheral device interface 1803 can be used to connect at least one I / O (Input / Output) related peripheral device to the processor 1801 and the memory 1802. In some embodiments, the processor 1801, the memory 1802, and the peripheral device interface 1803 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 1801, the memory 1802, and the peripheral device interface 1803 can be implemented on a separate chip or circuit board, and this embodiment does not limit this.

[0315] The radio frequency circuit 1804 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The radio frequency circuit 1804 communicates with a communication network and other communication devices through electromagnetic signals. The radio frequency circuit 1804 converts an electrical signal into an electromagnetic signal for transmission, or converts a received electromagnetic signal into an electrical signal. Optionally, the radio frequency circuit 1804 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, and so on. The radio frequency circuit 1804 can communicate with other terminals through at least one wireless communication protocol. The wireless communication protocol includes but is not limited to: the World Wide Web, a metropolitan area network, an intranet, each generation of mobile communication network (2G, 3G, 4G, and 5G), a wireless local area network, and / or a WiFi (Wireless Fidelity) network. In some embodiments, the radio frequency circuit 1804 may further include a circuit related to NFC (Near Field Communication), and this application does not limit this.

[0316] The display screen 1805 is used to display the UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof. When the display screen 1805 is a touch display screen, the display screen 1805 also has the ability to collect touch signals on or above the surface of the display screen 1805. The touch signal can be input as a control signal to the processor 1801 for processing. At this time, the display screen 1805 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard. In some embodiments, there may be one display screen 1805, which is disposed on the front panel of the computer device 1800; in other embodiments, there may be at least two display screens 1805, which are respectively disposed on different surfaces of the computer device 1800 or are in a folded design; in other embodiments, the display screen 1805 may be a flexible display screen, which is disposed on a curved surface or a folding surface of the computer device 1800. Even further, the display screen 1805 can also be set to an irregular non-rectangular shape, that is, a special-shaped screen. The display screen 1805 can be prepared using materials such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode).

[0317] The camera module 1806 is used to collect images or videos. Optionally, the camera module 1806 includes a front camera and a rear camera. Generally, the front camera is disposed on the front panel of the terminal, and the rear camera is disposed on the back of the terminal. In some embodiments, there are at least two rear cameras, which are any one of a main camera, a depth-of-field camera, a wide-angle camera, and a telephoto camera, to implement functions such as background blurring by fusing the main camera and the depth-of-field camera, panoramic shooting by fusing the main camera and the wide-angle camera, and VR (Virtual Reality) shooting function or other fusion shooting functions. In some embodiments, the camera module 1806 may further include a flash. The flash can be a single-color temperature flash or a two-color temperature flash. A two-color temperature flash refers to a combination of a warm light flash and a cold light flash, which can be used for light compensation under different color temperatures.

[0318] The audio circuit 1807 may include a microphone and a speaker. The microphone is used to collect sound waves of the user and the environment, and convert the sound waves into electrical signals, which are input to the processor 1801 for processing, or input to the radio frequency circuit 1804 to implement voice communication. For the purpose of stereo collection or noise reduction, there may be multiple microphones, which are respectively arranged at different parts of the computer device 1800. The microphone may also be an array microphone or an omnidirectional collection microphone. The speaker is used to convert the electrical signal from the processor 1801 or the radio frequency circuit 1804 into sound waves. The speaker may be a traditional thin film speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can not only convert the electrical signal into sound waves audible to humans, but also convert the electrical signal into sound waves inaudible to humans for uses such as ranging. In some embodiments, the audio circuit 1807 may further include a headphone jack.

[0319] The positioning component 1808 is used to locate the current geographical location of the computer device 1800 to implement navigation or LBS (Location Based Service). The positioning component 1808 may be a positioning component based on the GPS (Global Positioning System) of the United States, the Beidou system of China, or the Galileo system of Russia.

[0320] The power supply 1809 is used to supply power to each component in the computer device 1800. The power supply 1809 may be alternating current, direct current, a disposable battery, or a rechargeable battery. When the power supply 1809 includes a rechargeable battery, the rechargeable battery may be a wired rechargeable battery or a wireless rechargeable battery. A wired rechargeable battery is a battery charged through a wired line, and a wireless rechargeable battery is a battery charged through a wireless coil. The rechargeable battery can also be used to support fast charging technology.

[0321] In some embodiments, the computer device 1800 further includes one or more sensors 1810. The one or more sensors 1810 include, but are not limited to: an acceleration sensor 1811, a gyroscope sensor 1812, a pressure sensor 1813, a fingerprint sensor 1814, an optical sensor 1815, and a proximity sensor 1816.

[0322] The acceleration sensor 1811 can detect the magnitude of acceleration on the three coordinate axes of the coordinate system established with the computer device 1800. For example, the acceleration sensor 1811 can be used to detect the components of the gravitational acceleration on the three coordinate axes. The processor 1801 can control the display screen 1805 to display the user interface in a landscape view or a portrait view according to the gravitational acceleration signal collected by the acceleration sensor 1811. The acceleration sensor 1811 can also be used for game or collection of the user's motion data.

[0323] The gyroscope sensor 1812 can detect the body orientation and rotation angle of the computer device 1800. The gyroscope sensor 1812 can cooperate with the acceleration sensor 1811 to collect the 3D actions of the user on the computer device 1800. Based on the data collected by the gyroscope sensor 1812, the processor 1801 can implement the following functions: motion sensing (such as changing the UI according to the user's tilting operation), image stabilization during shooting, game control, and inertial navigation.

[0324] The pressure sensor 1813 can be disposed on the side frame of the computer device 1800 and / or the lower layer of the display screen 1805. When the pressure sensor 1813 is disposed on the side frame of the computer device 1800, it can detect the holding signal of the user on the computer device 1800, and the processor 1801 can identify the left and right hands or perform a quick operation according to the holding signal collected by the pressure sensor 1813. When the pressure sensor 1813 is disposed on the lower layer of the display screen 1805, the processor 1801 can control the operable controls on the UI interface according to the pressure operation of the user on the display screen 1805. The operable controls include at least one of a button control, a scroll bar control, an icon control, and a menu control.

[0325] The fingerprint sensor 1814 is used to collect the fingerprints of the user. The processor 1801 can identify the user's identity according to the fingerprints collected by the fingerprint sensor 1814, or the fingerprint sensor 1814 can identify the user's identity according to the collected fingerprints. When the identified user identity is a trusted identity, the processor 1801 authorizes the user to perform relevant sensitive operations, and the sensitive operations include unlocking the screen, viewing encrypted information, downloading software, making payments, and changing settings, etc. The fingerprint sensor 1814 can be disposed on the front, back, or side of the computer device 1800. When there are physical buttons or a manufacturer logo on the computer device 1800, the fingerprint sensor 1814 can be integrated with the physical buttons or the manufacturer logo.

[0326] The optical sensor 1815 is used to collect the ambient light intensity. In one embodiment, the processor 1801 can control the display brightness of the display screen 1805 according to the ambient light intensity collected by the optical sensor 1815. Specifically, when the ambient light intensity is high, the display brightness of the display screen 1805 is increased; when the ambient light intensity is low, the display brightness of the display screen 1805 is decreased. In another embodiment, the processor 1801 can also dynamically adjust the shooting parameters of the camera module 1806 according to the ambient light intensity collected by the optical sensor 1815.

[0327] A proximity sensor 1816, also known as a distance sensor, is typically disposed on the front panel of the computer device 1800. The proximity sensor 1816 is used to collect the distance between the user and the front of the computer device 1800. In one embodiment, when the proximity sensor 1816 detects that the distance between the user and the front of the computer device 1800 is gradually decreasing, the processor 1801 controls the display screen 1805 to switch from the lit state to the off state; when the proximity sensor 1816 detects that the distance between the user and the front of the computer device 1800 is gradually increasing, the processor 1801 controls the display screen 1805 to switch from the off state to the lit state.

[0328] Those skilled in the art can understand that Figure 18 the structure shown in does not constitute a limitation on the computer device 1800, and may include more or fewer components than shown, or combine certain components, or adopt a different component arrangement.

[0329] This application also provides a computer-readable storage medium, in which at least one instruction, at least one segment of program, code set or instruction set is stored, and the at least one instruction, the at least one segment of program, the code set or instruction set is loaded and executed by a processor to implement the infectious disease prediction method and the training method of the graph convolutional neural network provided by the above method embodiments.

[0330] This application provides a computer program product or a computer program, the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the infectious disease prediction method and the training method of the graph convolutional neural network provided by the above method embodiments.

[0331] The serial numbers of the above embodiments of this application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0332] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above embodiments can be completed by hardware, or can be completed by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disc, etc.

[0333] The above are only optional embodiments of this application and are not intended to limit this application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this application shall be included in the protection scope of this application.

Claims

1. A method for predicting an infectious disease, characterized in that, The method includes: Obtain the first physiological information and the first location information of the first user, where the first physiological information is the physiological information of the first user related to the infectious disease, and the first location information indicates the location where the first user is located; Based on the first physiological information, generate a first user node in the basic graph network of the infectious disease; the basic graph network includes a first region node, and the first region node identifies the first region where the user suspected of having the infectious disease is located; When the location where the first user is located indicated by the first location information falls within the first region, perform an edge connection operation between the first user node and the first region node in the basic graph network to form a first graph network; Input the first graph network into a graph convolutional neural network, and predict to obtain a first user feature vector of the first user node; and, based on the first user feature vector, predict the probability that the first user has the infectious disease through a probability calculation network.

2. The method according to claim 1, wherein The graph convolutional neural network includes k graph convolutional layers; The step of inputting the first graph network into a graph convolutional neural network and predicting to obtain a first user feature vector of the first user node includes: Based on the historical disease information carried by the first region node in the first graph network, obtain an initial region feature vector; based on the first physiological information carried by the first user node in the first graph network, obtain an initial user feature vector; Based on the initial region feature vector and the initial user feature vector, calculate to obtain a first-layer region feature vector; based on the initial region feature vector and the initial user feature vector, calculate to obtain a first-layer user feature vector; Based on the (i - 1)-th layer region feature vector and the (i - 1)-th layer user feature vector, calculate to obtain the i-th layer region feature vector; based on the (i - 1)-th layer region feature vector and the (i - 1)-th layer user feature vector, calculate to obtain the i-th layer user feature vector; Output the k-th layer user feature vector as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

3. A method for predicting an infectious disease, characterized in that, The method includes: Obtain the first physiological information and the first location information of the first user, where the first physiological information is the physiological information of the first user related to the infectious disease, and the first location information indicates the location where the first user is located; Based on the first physiological information, generate a first user node in the basic graph network of the infectious disease; the basic graph network includes a first region node and a second region node connected to the first region node, the first region node identifies the first region where the user suspected of having the infectious disease is located, and the second region node identifies the second region where the user suspected of having the infectious disease is located; When the location where the first user is located indicated by the first location information falls within the first region, perform an edge connection operation between the first user node and the first region node in the basic graph network to form a second graph network; Wherein, the number of population flows between the second regional node and the first regional node reaches a quantity threshold, and / or, the distance between the regions indicated by the second regional node and the first regional node is lower than a first distance threshold; Input the second graph network into a graph convolutional neural network to predict a first user feature vector of the first user node; and, based on the first user feature vector, predict the probability that the first user has the infectious disease through a probability calculation network.

4. The method according to claim 3, wherein The graph convolutional neural network includes k graph convolutional layers; The inputting the second graph network into a graph convolutional neural network to predict a first user feature vector of the first user node includes: Based on the historical disease information carried by the first regional node in the second graph network, obtain an initial regional feature vector of the first regional node; based on the historical disease information carried by the second regional node in the second graph network, obtain an initial regional feature vector of the second regional node; based on the first physiological information carried by the first user node in the second graph network, obtain an initial user feature vector; Based on the initial regional feature vector of the first regional node, the initial regional feature vector of the second regional node, and the initial user feature vector, combined with a first weight, calculate a first-layer regional feature vector of the first regional node, where the first weight is associated with the number of population flows between the first regional node and the second regional node; based on the initial regional feature vector of the first regional node and the initial user feature vector, calculate a first-layer user feature vector; Based on the (i-1)-th layer regional feature vector of the first regional node, the (i-1)-th layer regional feature vector of the second regional node, and the (i-1)-th layer user feature vector, combined with the first weight, calculate an i-th layer regional feature vector of the first regional node; based on the (i-1)-th layer regional feature vector of the first regional node and the (i-1)-th layer user feature vector, calculate an i-th layer user feature vector; Output the k-th layer user feature vector as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

5. A method for predicting an infectious disease, characterized in that, The method includes: Obtain first physiological information and first location information of a first user, where the first physiological information is physiological information related to the infectious disease of the first user, and the first location information indicates the location where the first user is located; Based on the first physiological information, generate a first user node in the basic graph network of the infectious disease; the basic graph network includes a first regional node and a second user node connected to the first regional node, where the first regional node identifies the first region where the user suspected of having the infectious disease is located, and the second user node identifies a second user suspected of having the infectious disease; In the case where the location where the first user is located indicated by the first location information falls within the first region, perform an edge connection operation between the first user node and the first regional node in the basic graph network to form a third graph network; Among them, the location where the second user indicated by the second user node falls within the first region; Input the third graph network into a graph convolutional neural network to predict the first user feature vector of the first user node; and, based on the first user feature vector, predict the probability that the first user has the infectious disease through a probability calculation network.

6. The method according to claim 5, wherein The graph convolutional neural network includes k graph convolutional layers; The inputting the third graph network into a graph convolutional neural network to predict the first user feature vector of the first user node includes: Based on the historical disease information carried by the first region node in the third graph network, obtain an initial region feature vector; based on the first physiological information carried by the first user node in the third graph network, obtain the initial user feature vector of the first user node; based on the second physiological information carried by the second user node in the third graph network, obtain the initial user feature vector of the second user node; Based on the initial region feature vector, the initial user feature vector of the first user node, and the initial user feature vector of the second user node, calculate the first-layer region feature vector; based on the initial region feature vector and the initial user feature vector of the first user node, calculate the first-layer user feature vector of the first user node; Based on the (i - 1)-th layer region feature vector, the (i - 1)-th layer user feature vector of the first user node, and the (i - 1)-th layer user feature vector of the second user node, calculate the i-th layer region feature vector; based on the (i - 1)-th layer region feature vector and the (i - 1)-th layer user feature vector of the first user node, calculate the i-th layer user feature vector of the first user node; Output the k-th layer user feature vector of the first user node as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

7. A method for predicting an infectious disease, characterized in that, The method includes: Obtain the first physiological information and the first location information of the first user, where the first physiological information is the physiological information related to the infectious disease of the first user, and the first location information indicates the location where the first user is located; Based on the first physiological information, generate a first user node in the basic graph network of the infectious disease; the basic graph network includes a third user node, and the third user node identifies a third user suspected of having the infectious disease, and the third user node carries the third physiological information and the third location information of the third user; In the case where the similarity between the first physiological information and the third physiological information is greater than the similarity threshold, and / or the distance between the locations indicated by the first location information and the third location information is less than the second distance threshold, perform an edge connection operation between the first user node and the third user node in the basic graph network to form a fourth graph network; Input the fourth graph network into a graph convolutional neural network to predict a first user feature vector of the first user node; and, based on the first user feature vector, predict the probability that the first user has the infectious disease through a probability calculation network.

8. The method according to claim 7, wherein The graph convolutional neural network includes k graph convolutional layers; The step of inputting the fourth graph network into a graph convolutional neural network to predict a first user feature vector of the first user node includes: Based on the first physiological information carried by the first user node in the fourth graph network, obtain an initial user feature vector of the first user node; based on the third physiological information carried by the third user node in the fourth graph network, obtain an initial user feature vector of the third user node; Based on the initial user feature vector of the first user node and the initial user feature vector of the third user node, and in combination with a second weight, calculate a first-layer user feature vector of the first user node and a first-layer user feature vector of the third user node, where the second weight is associated with at least one of the similarity between the first physiological information and the third physiological information and the distance between the positions indicated by the first location information and the third location information; Based on the (i - 1)-th layer user feature vector of the first user node and the (i - 1)-th layer user feature vector of the third user node, and in combination with the second weight, calculate an i-th layer user feature vector of the first user node and an i-th layer user feature vector of the third user node; Output the k-th layer user feature vector of the first user node as the first user feature vector, where k is an integer greater than 1, and i is an integer greater than 0 and less than k.

9. The method according to claim 1 or 3 or 5 or 7, characterized in that, The probability calculation network includes an activation function; The step of predicting the probability that the first user has the infectious disease through a probability calculation network based on the first user feature vector includes: Concatenate the first user feature vector and the initial user feature vector of the first user node to obtain a first intermediate vector; Input the first intermediate vector into a fully connected network to obtain a second intermediate vector; Input the second intermediate vector into the activation function to predict the probability that the first user has the infectious disease.

10. The method according to claim 1 or 3 or 5 or 7, characterized in that For the training method of the graph convolutional neural network, the training method adopts a semi-supervised mode, and the method includes: Obtain sample physiological information and sample location information of n sample users; the n sample users include n1 first sample users and n2 second sample users, the first sample users carry labels for determining whether the first sample users have the infectious disease, and the second sample users do not carry the labels; Input the sample graph network corresponding to the sample physiological information and sample location information of the n sample users into the graph convolutional neural network, and calculate n sample feature vectors of the n sample users through the graph convolutional neural network; For one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has the infectious disease through the probability calculation network; Based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users, train the graph convolutional neural network.

11. An infectious disease prediction device, characterized in that, The device is used to implement the infectious disease prediction method as described in claim 1 or claim 3 or claim 5 or claim 7. The device includes: An acquisition module, configured to acquire the first physiological information and the first location information of a first user, where the first physiological information is the physiological information related to the infectious disease of the first user, and the first location information indicates the location where the first user is located; An update module, configured to add a first user node to the basic graph network of the infectious disease based on the first physiological information and the first location information to obtain an updated graph network; the basic graph network includes at least one user node and / or at least one area node, the user node identifies a user suspected of having the infectious disease, and the area node identifies the area where the user suspected of having the infectious disease is located; A prediction module, configured to input the updated graph network into a graph convolutional neural network to predict and obtain a first user feature vector of the first user; and based on the first user feature vector, predict the probability that the first user has the infectious disease through a probability calculation network.

12. The device according to claim 11, characterized in that The device is also used to implement the graph convolutional neural network. The device includes: An acquisition module, configured to acquire the sample physiological information and sample location information of n sample users; the n sample users include n1 first sample users and n2 second sample users, the first sample users carry labels, and the labels are used to determine whether the first sample users have the infectious disease, and the second sample users do not carry the labels; A calculation module, configured to input the sample physiological information and sample location information of the n sample users into the graph convolutional neural network, and calculate and obtain n sample feature vectors of the n sample users through the graph convolutional neural network; The calculation module is further configured to, for one of the n1 first sample users, based on the sample feature vector of the first sample user, obtain the predicted probability that the first sample user has the infectious disease through the probability calculation network; A training module, configured to train the graph convolutional neural network based on the loss value between the predicted probability and the proportion of diseased sample users among the n1 first sample users.

13. A computer device, characterized in that, The computer device includes: a processor and a memory. The memory stores a computer program, and the computer program is loaded and executed by the processor to implement the infectious disease prediction method as described in any one of claims 1 to 10.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and the computer program is loaded and executed by a processor to implement the infectious disease prediction method as described in any one of claims 1 to 10.

15. A computer program product, characterized in that, The computer program product includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the infectious disease prediction method according to any one of claims 1 to 10.

Images